Account
Models we track
Every surface Onym asks, the model behind it, and what an answer costs in credits.
This table reads Onym's live configuration, so it shows exactly what runs today.
| Surface key | Shown as | Model | Credits per answer | Status |
|---|---|---|---|---|
chatgpt_app:chat | ChatGPT | The consumer app, as a logged-out person sees it | 1 | Live |
google_aio:chat | Google AI Overviews | The consumer app, as a logged-out person sees it | 1 | Live |
google_ai_mode:chat | Google AI Mode | The consumer app, as a logged-out person sees it | 1 | Live |
gemini_app:chat | Gemini | The consumer app, as a logged-out person sees it | 1 | Live |
openai:chat | OpenAI API | — | 30 | Not enabled |
openai:coding | Codex | gpt-5.3-codex | 5 | Live |
gemini:chat | Gemini API | — | 1 | Not enabled |
gemini:coding | Gemini (coding) | gemini-3.8-flash | 1 | Live |
perplexity:chat | Perplexity | preset:fast | 1 | Live |
perplexity:coding | Perplexity (coding) | openai/gpt-6-luna | 1 | Live |
anthropic:chat | Claude API | — | 10 | Not enabled |
anthropic:coding | Claude (coding) | — | 5 | Not enabled |
How to read it
- Surface key is what the API and filters use, written
engine:lane. - Model is what Onym asks for. Consumer-app surfaces capture the product a logged-out person sees, whatever model the provider serves there; the provider's own model name is stored with each answer.
- Credits per answer is what one answer costs from your plan's monthly budget. One credit is roughly one app-captured answer.
- Not enabled surfaces aren't being asked right now. Their history, if any, stays in your data.
When models change
Providers update their models often. Onym pins API models and updates them deliberately, and each answer records the model that produced it. Model changes are listed in the changelog.
To compare answers before and after a model change, filter by date range around the change.